Scribd
Upload
105 views

The Riemann Zeta Function

The Riemann zeta function ζ(p) is defined as the sum of the reciprocals of natural numbers raised to the power of p, which converges for p > 1. For p = 2, the function equals π2/6. Computing this value to within an error of 10-6 using the direct sum requires summing over 1 million terms. However, using the alternating series definition of the zeta function f(p) improves the rate of convergence, requiring only 1000 terms to achieve the same accuracy. This alternating series definition of the zeta function can be extended to all values of p > 0, suggesting the domain of the Riemann zeta function could also be extended in this way.

Uploaded by

নিয়ৰ শৰ্মা
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
105 views

The Riemann Zeta Function

The Riemann zeta function ζ(p) is defined as the sum of the reciprocals of natural numbers raised to the power of p, which converges for p > 1. For p = 2, the function equals π2/6. Computing this value to within an error of 10-6 using the direct sum requires summing over 1 million terms. However, using the alternating series definition of the zeta function f(p) improves the rate of convergence, requiring only 1000 terms to achieve the same accuracy. This alternating series definition of the zeta function can be extended to all values of p > 0, suggesting the domain of the Riemann zeta function could also be extended in this way.

Uploaded by

নিয়ৰ শৰ্মা
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

The Riemann Zeta Function

The Riemann zeta function is dened by the p-series (p) = 1 1 1 1 = 1 + p + p + p + ..., p 2 3 4 n=1 n

valid for p > 1,

(1)

which converges for p > 1 by the Integral Test (and diverges for p 1). One interesting special value [though hard to prove] is (2) = 1 1 1 1 1 2 = 1 + 2 + 2 + 2 + 2 + ... = . 2 2 3 4 5 6 n=1 n

(2)

Convergence Suppose we try to use equation (2) to compute 2 /6 with an error of less than 106 . Question: How many terms do we need? Explicitly, how large must k be to bring k 1/n2 within 106 of 2 /6? n=1 We note rst that 1/n2 < 106 whenever n > 1000. This suggests taking k = 1000. But consider the next thousand terms, each of which is at least 1/20002 ; their sum is therefore greater than 1000 1/20002 = 0.000250, which is far larger than 106 . [The error is actually 0.0009995, according to maple. In fact, maple knows about the series equation (1), and the remainder after k terms is a built-in function.] The correct answer is (exactly) 1,000,000, with error 9.999995 107 . Thus equation (1) is highly impractical for computation. An alternating series We introduce the variant

f (p) =
n=1

(1)n1

1 1 1 1 = 1 p + p p + ..., p n 2 3 4

valid for p > 0,

(3)

which converges for p > 0 by the Alternating Series Test. Rearrangement The function f (p) is easily expressed in terms of (p), when p > 1. Consider the odd and even terms separately, by dening g(p) = 1 + and h(p) = 1 1 1 + p + p + ..., p 2 4 6 valid for p > 1. 1 1 1 + p + p + ..., p 3 5 7 valid for p > 1,

Clearly, g(p)+h(p) = (p). This rearrangement is valid for p > 1. Also, multiplication of each term of h(p) by 2p gives 2p h(p) = (p). We have enough information to deduce that h(p) = Then f (p) = g(p) h(p) gives f (p) = 110.109 Calculus II 2p 1 1 2p 2 p (p) = (p) . 2p 2 2p (5) 1 (p), 2p g(p) = 2p 1 (p) . 2p (4)

JMB File: zeta, Revision A; 30 Aug 2001; Page 1

The Riemann Zeta Function

More examples For p = 2, equation (4) gives two more well-known series, g(2) = 1 + and f (2) = 1 1 1 1 3 2 + 2 + 2 + . . . = (2) = 32 5 7 4 8 1 1 1 2 + 2 2 + ... = . 22 3 4 12 (6)

(7)

By Theorem 8, we only need 1000 terms of the series equation (7) to guarantee that the error is less than 106 . [The actual error is only 4.997 107 .] This is a big improvement over equation (2), but still hardly practical. Extension Nevertheless, equation (3) is useful for more than just computation. We can turn equation (5) around to the form (p) = 2p f (p) for p > 1. 2p 2 (8)

However, the right side of this equation is dened for all p > 0 (provided we exclude the case p = 1, so as not to divide by 0). This suggests that it is reasonable to extend the denition of (p) to all p > 0 (p = 1) by means of this equation. [This is in fact true, but requires more justication than this.]

110.109 Calculus II

JMB File: zeta, Revision A; 30 Aug 2001; Page 2

You might also like